Detecting abandoned packages in a multi-camera video surveillance system

被引：33

作者：

Beynon, MD ^{[1
]}

Van Hook, DJ ^{[1
]}

Seibert, M ^{[1
]}

Peacock, A ^{[1
]}

Dudgeon, D ^{[1
]}

机构：

[1] MIT, Lincoln Lab, Lexington, MA 02173 USA

来源：

IEEE CONFERENCE ON ADVANCED VIDEO AND SIGNAL BASED SURVEILLANCE, PROCEEDINGS | 2003年

关键词：

D O I：

10.1109/AVSS.2003.1217925

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We describe a video surveillance system that detects abandoned packages automatically. In this system, multiple cameras locate objects in space and time despite occlusions and distracting lighting effects observed by subsets of the cameras. A multiple-state model of an abandoned package provides the ability to detect realistic abandoned package events. This paper outlines the system by describing the modules for camera view segmentation, object classification, view-object association, 3D object tracking, and finally detection of the event of a package being abandoned, which highlights object association using multiple cameras and the multiple-state abandoned package model.

引用

页码：221 / 228

页数：8

共 17 条

[1]

[Anonymous], 2000, 3 IEEE INT WORKSH VI

[2]

[Anonymous], TR991730 CORN U DEP

[3]

BOWLES J, 1998, P SPIE C INFR SPAC R, V6, P20

[4] Multiple camera fusion for multi-object tracking [J].

Dockstader, SL ;

Tekalp, AM .

2001 IEEE WORKSHOP ON MULTI-OBJECT TRACKING, PROCEEDINGS, 2001, :95-102

[5] Detecting suspicious background changes in video surveillance of busy scenes [J].

Gibbins, D ;

Newsam, GN ;

Brooks, MJ .

THIRD IEEE WORKSHOP ON APPLICATIONS OF COMPUTER VISION - WACV '96, PROCEEDINGS, 1996, :22-26

[6]

Hart, 2006, PATTERN CLASSIFICATI

[7]

Harville M., 2001, P IEEE WORKSH DET RE

[8]

Horprasert T., 1999, P IEEE ICCV 99 FRAME

[9] Recognizing planned, multiperson action [J].

Intille, SS ;

Bobick, AF .

COMPUTER VISION AND IMAGE UNDERSTANDING, 2001, 81 (03) :414-445

[10]

JOHNSON DH, 1993, ARRAY SIGNAL PROCESS, pCH8

← 1 2 →